<h1 id="premium-character-loras">Premium character LoRAs</h1>Original synthetic characters, bundles and AI tools are in my shop:<a target="_blank" rel="ugc" href="https://rkrra.com/civitai">https://rkrra.com/civitai</a>Get the character LoRAs on Patreon: <a target="_blank" rel="ugc" href="https://www.patreon.com/cw/RkRRa">https://www.patreon.com/cw/RkRRa</a>Join the community on Discord: <a target="_blank" rel="ugc" href="https://discord.gg/nB79wRNtMT">https://discord.gg/nB79wRNtMT</a><h2>Grotesk Female 1</h2>A character LoRA for Krea 2. Reproduces a consistent subject across poses, framing, and lighting.<h3>Usage</h3>Include the trigger word <code>grtskfml1</code> in your prompt.Recommended weight: 0.7. At weight 1 this LoRA also pulls other people in the image towards this character: ask for someone else and you may still get this face. At 0.7 the likeness holds and the rest of your prompt is respected. If you need a stronger likeness, go up in small steps.Two character LoRAs at once tend to blend, especially when the two characters look alike. For two characters in one image, generate them one at a time.<h3>Prompt Examples</h3>Here are some prompt examples:<blockquote>"grtskfml1, chest-up, eye-level, facing right, lips slightly parted, snowy village with wooden buildings, mountains in distance, soft golden hour light, quiet street."</blockquote><blockquote>"grtskfml1, chest-up, straight-on, standing still, neutral expression, night market stalls blurred behind, bright string lights overhead, soft ambient glow from surrounding booths, shallow depth of field."</blockquote><blockquote>"grtskfml1, full length from behind, straight-on, standing centered, facing away, bookshelves lining both sides, bright window ahead, even illumination, focused posture, elongated corridor perspective."</blockquote><blockquote>"grtskfml1, chest-up, slightly low angle, head turned right, lips parted, neutral expression, plain warm-toned wall, soft directional light from left, highlights contours, shadows fall right, minimal background distraction, focused on upper body."</blockquote><blockquote>"grtskfml1, chest-up, straight-on, arms crossed, neutral expression, snowy mountain village backdrop, bright daylight, cobblestone path, wooden chalets, snow-covered roofs, distant peaks, clear sky."</blockquote><blockquote>"grtskfml1, three-quarter from mid-thigh up, slightly low angle, seated cross-legged on floor, hands clasped, head tilted down, olive jacket over grey tee, dark pants, bookshelf and sofa in background, soft natural light from side."</blockquote><blockquote>"grtskfml1, chest-up, slightly low angle, head tilted back, mouth open wide, laughing, dark plain background, soft directional lighting from front-left, highlighting contours, creating subtle shadows, emphasizing form and texture."</blockquote><blockquote>"grtskfml1, close-up portrait, eye-level, facing forward, neutral expression, gray coat over turtleneck, urban street background, soft diffused daylight."</blockquote><blockquote>"grtskfml1, three-quarter from mid-thigh up, straight-on, seated on bench, arms resting on knees, neutral expression, park with trees, warm golden hour light filtering through foliage."</blockquote><blockquote>"grtskfml1, full length from behind, straight-on, seated on table, facing window, back to viewer, train interior, natural daylight from window, soft ambient light, minimal movement."</blockquote><blockquote>"grtskfml1, full length from behind, straight-on, hands on hips, back turned, teal wall backdrop, bright even lighting, clean minimal background, confident stance, no facial expression visible, vibrant color contrast, smooth surface reflections."</blockquote><blockquote>"grtskfml1, three-quarter from mid-thigh up, slightly low angle, arms raised high, hands open, joyful expression, urban skyline backdrop, overcast daylight, rooftop setting, concrete surface, distant buildings, elevated vantage point."</blockquote><blockquote>"grtskfml1, chest-up, straight-on, standing, mouth open, expressive, plain blue background, even studio lighting, direct gaze, neutral backdrop, focused on upper body, clear visibility."</blockquote><blockquote>"grtskfml1, chest-up, straight-on, arms crossed, stern gaze, plain terracotta wall, soft even lighting, direct eye contact, confident stance, neutral backdrop, controlled studio environment."</blockquote><blockquote>"grtskfml1, chest-up, straight-on, facing camera, lips puckered, teal backdrop, even lighting, stylized makeup, spiky hair, pointed ears, hoop earrings, metallic collar, pink top."</blockquote><blockquote>"grtskfml1, waist-up, straight-on, hands on hips, neutral expression, black leather jacket, dark jeans, teal backdrop, even studio lighting."</blockquote><blockquote>"grtskfml1, three-quarter from mid-thigh up, straight-on, standing on beach, looking forward, ocean behind, warm golden hour light, sand beneath feet, arms slightly away from body, confident stance."</blockquote><blockquote>"grtskfml1, three-quarter from mid-thigh up, slightly low angle, turning back over shoulder, smiling broadly, inside train carriage with rows of seats, natural light from windows, soft shadows on floor and seats."</blockquote><blockquote>"grtskfml1, chest-up, slightly low angle, turned head, looking sideways, neutral expression, plain gray background, soft studio lighting, even illumination, minimal shadows, focused on upper body, clean backdrop."</blockquote><blockquote>"grtskfml1, chest-up, straight-on, facing camera, mouth open, surprised expression, kitchen background, warm overhead lighting, wooden cabinets, tiled backsplash, soft shadows, indoor setting."</blockquote><h3>Training</h3><ul><li>Base model: Krea 2 (trained on Raw)</li><li>Type: Character LoRA</li><li>Dataset: 46 images</li><li>Steps: 2000</li><li>Network rank / alpha: 32 / 32</li><li>Learning rate: 0.0002</li><li>Optimizer: adamw8bit</li><li>Resolutions: 512 / 1024</li></ul><h3>Generation settings</h3>Trained on Krea 2 Raw, and transfers to Krea 2 Turbo for faster generation.<ul><li>Raw: ~30 steps, guidance 3.5</li><li>Turbo: ~8 steps, guidance 1.0</li></ul>The preview images were generated with <code>krea2_turbo_mxfp8.safetensors</code>, at 8 steps, guidance 1, LoRA weight 1.

Files with SHA256: e69f6d408f202bc68a77c82f517460a94f373aaf842d29906ae97ecec7d2ece4